Drying Kinetics and Morphology - American Chemical Society

Feb 17, 2009 - Department of Chemical Engineering, Technical UniVersity of Denmark, Building 229, DK-2800 Kgs. Lyngby,. Denmark, and Solid Products De...
9 downloads 10 Views 2MB Size
Ind. Eng. Chem. Res. 2009, 48, 3657–3664

3657

Spray Drying of Suspensions for Pharma and Bio Products: Drying Kinetics and Morphology Jakob Sloth,†,‡ Kåre Jørgensen,§ Poul Bach,‡ Anker D. Jensen,*,‡ Søren Kiil,‡ and Kim Dam-Johansen‡ Department of Chemical Engineering, Technical UniVersity of Denmark, Building 229, DK-2800 Kgs. Lyngby, Denmark, and Solid Products DeVelopment, NoVozymes A/S, KrogshøjVej 36, DK-2880 BagsVærd, Denmark

An experimental investigation of the spray drying behavior of droplets containing excipients and carrier materials used in the pharmaceutical and biotechnological industries has been conducted. Specifically, rice starch suspensions with different amounts of TiO2, maltodextrin, dextrin, NaCl and Na2SO4 are dried. The drying rate is measured, and the morphology formation is mapped to obtain a more fundamental understanding of the drying process, which is very useful when designing product formulations. In the pilot spray dryer, droplet generation is based on the JetCutter technology and the droplets are dried under well-defined temperature and flow conditions. The droplets are sampled during drying to determine the drying rate, and the dried particles are collected for morphology analysis. The results show that reducing the water activity in a suspension of insolubles by adding various amounts of inorganic salts or carbohydrates causes an increase in the droplet temperature during spray drying resulting in a rather constant the drying rate. Further, the results show that small alterations in the droplet composition may significantly change the final particle morphology. The observed morphologies are discussed in detail. Introduction Spray drying is commonly used in the manufacture of biotechnological and pharmaceutical products. The products are often of high value, and it is therefore important that they are produced with a constant and high quality. Consequently, the choices of formulation ingredients, equipment design, and process conditions become critical and preferably rely on a detailed insight into the various processes occurring during product drying. In spray drying, a suspension or solution is fed to an atomizer and the droplets formed are mixed with a hot gas. This causes the solvent of the droplets to evaporate, leading to formation of particulates. The course of drying for a single droplet inside a spray dryer is complex, involving stages of different droplet temperatures and evaporation rates. The process is described in detail by several authors (see, e.g., the work of Nesic and Vodnik,1 Farid,2 or Sloth et al.3) and progresses as shown in Figure 1. The figure is valid both for the drying of a droplet consisting of a solution or a suspension. A suspension is a mixture of solvent and primary particles, i.e. insoluble particles which normally have a size of than less 10 µm. The following briefly describes drying of a suspension droplet. After being formed at the atomizing device, the droplet is heated (stage A) and the temperature approaches the wet bulb temperature of the solvent. During stage B, the surface is fully wetted by the solvent, giving rise to fast evaporation. Due to the evaporation, the size of the droplet decreases and the primary particle concentration builds up, inducing minor droplet heating because of slight solvent evaporation hindrance. When the primary particles at the droplet surface have packed as closely as possible (stage C), a solid phase forms around the wet core of the droplet. The solid phase limits the water transport * To whom correspondence should be addressed. Tel.: +45 45 25 28 41. Fax: +45 45 88 22 58. E-mail: [email protected]. † Present address: Niro A/S, Gladsaxevej 305, 2860 Søborg, Denmark. ‡ Technical University of Denmark. § Solid Products Development.

from the wet core to the surface and in the remaining part of the drying process the rate of evaporation falls while the temperature approaches the dry bulb temperature (stage D). Throughout the entire drying process, there is a strong coupling between the droplet temperature and the drying rate. If a decrease in the drying rate occurs, less heat is used for evaporation and the droplet temperature increases. A higher temperature induces accelerated drying, lowering the temperature. However, an equilibrium between temperature and drying rate is quickly obtained. The description above is widely accepted in the literature where stage B is often referred to as the constant rate period, and stages C and D are combined and referred to as the falling rate period. Previous Studies The development in the droplet mass shown in Figure 1 is merely qualitative. The specific drying rate of a droplet depends on numerous variables such as composition, initial droplet size, drying air temperature, and humidity. However, due to the complex flow field in a spray dryer, measuring the drying rate is hardly possible. Droplets released from the atomizer have different velocities and directions, and they travel through regions of varying temperatures and humidities. It is this complexity which renders the experimental investigations of drying kinetics almost impossible using a conventional industrial or pilot spray dryer. As a consequence, authors studying droplet drying kinetics resort to purpose-built equipment. In general, equipment of this kind may be divided into two groupsssingle droplet and in situ drying equipment. The former is usually performed by suspending a droplet from a filament. The stationary droplet is placed in an air stream with a constant temperature which may be well above 200 °C.4 The drying rate is measured as a mass loss by attaching the filament to a microbalance5,6 or a cathetometer.7 Alternatives include determining the deflection of a curved filament8 or suspending

10.1021/ie800983w CCC: $40.75  2009 American Chemical Society Published on Web 02/17/2009

3658 Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009

measuring up to 18 m.22 The droplets dry as they fall freely down the column in a heated air stream. Droplets are sampled along the height of the tower and subsequently subjected to moisture content analysis. The analysis may be vacuum or oven drying22,24 or Karl Fischer titration.23 The droplet sizes used during in situ experiments are in the same order of magnitude as that of industrial spray drying. The drying process inside the column resembles spray drying well22-25 and thus in situ experiments are well-suited both as complement to single droplet experiments and also as basis for new studies of drying kinetics and morphology. Figure 1. Progress of droplet drying in a spray dryer. (A) Heating. (B) Constant rate period. (C and D) Falling rate period.

the droplet from a burette and measuring the feed rate of water through the burette necessary to maintain a constant droplet diameter.9,10 The latter is obviously only possible for a pure liquid droplet. The approaches described above yield precise measurements of the drying rate throughout the entire course of drying and another advantage of the suspended droplet method is that it is possible to follow the progress in temperature by inserting a small thermocouple into the droplet.4,5,7,11 Also, the particle morphology formation may be monitored using a microscope camera.4,5,11 For practical reasons, the authors mentioned above are compelled to conduct experiments with initial droplet sizes in the range 0.8-2 mm. This is about an order of magnitude larger than spray droplets,12 and therefore, drying kinetics and resulting particle morphologies may not accurately reflect those in industrial spray drying equipment. Further, Lin and Gentry13 note that the filament prevents the droplet from rotating freely, serves as a heat source, and may deform the droplet when a solid phase forms in the later part of the drying process. Some of the difficulties associated with the filament are avoided when using ultrasonic levitation. Pioneering work on this technique is described by Toei et al.14 who successfully dried nonsupported droplets, held constant against gravity by the force of a standing acoustic wave. The drying rate during the constant rate period is easily measured using a camera as the droplet size changes. In contrast, the drying rate during the falling rate period is difficult to measure but enclosing the levitator in a preconditioned chamber and using a high accuracy dew point hygrometer to determine the amount of water evaporated was tested by Groenewold et al.15 Also, the vertical displacement of the droplet in the ultrasonic field is a function of the mass which, in theory, may be used for drying kinetics measurements.16,17 Droplet surface temperature measurements can be performed using by an infrared camera.18 A disadvantage of the levitation method is that if the droplets are too large (i.e., larger than 1.5 mm), the acoustic waves distort the shape of the droplets19 and therefore the experiments do not fully mimic the drying behavior in a spray dryer.20 Single droplet drying investigations have given excellent information on drying rates and droplet temperatures. The experiments map various phenomena of droplet drying, serve as validation of modeling studies, and thereby undoubtedly lead to a deeper understanding of the spray drying process. In a fairly recent review paper, Chen21 emphasizes the importance of in situ experiments as they closely imitate the evaporation process in conventional spray drying equipment. In the literature, a few authors describe the use of in situ experiments.22-25 Generally, the experimental approach is to situate an atomizing device at the top of a vertical drying column

Specific Objectives The objective of this work is to use in situ experimental investigations to contribute to the fundamental understanding of the droplet drying process taking place inside a spray dryer. The focus is on drying of formulations containing carrier materials and excipients often used in the biotechnological and pharmaceutical industries, omitting the active ingredient typically present in very low levels. Drying of these formulations is of particular interest as the influence on the drying process of each formulation ingredient considered may be mapped without the hazards of working with active components. Also, it is the objective of this paper to present an apparatus for in situ experimental investigations of drying droplets. Using this apparatus, the influence of formulation ingredients on the droplet drying kinetics and particle morphology formation are analyzed and discussed in detail. It is an important part of this work to connect the findings on morphology formation to the findings of other authors as this will contribute to the ongoing effort in the literature to understand morphology formation processes. Droplet Dryer An experimental apparatus called the droplet dryer which can be used to investigate the drying kinetics of multicomponent suspensions has been constructed. The design, construction, and commissioning process is described by Jørgensen.26 The droplet dryer meets the following requirements: • Droplets of an arbitrary size within the range 200-1500 µm may be produced. • The suspension feed may have a high initial solid content. • The drying takes place under well-defined conditions. • Sampling of the droplets during drying is possible. • The drying air temperature may be up to 250 °C. • The dried particles may be collected for, e.g., morphology analysis. Inspiration for the design of the droplet dryer has been drawn from the works of Bu¨ttiker,22 Wallack et al.,23 Meerdink,24 and Zbicinski.25 The layout of the setup is shown in Figure 2 while key specifications are given in Table 1. The apparatus consists of two partss droplet generator and a drying tower. These are described in detail below but the main principle of the droplet dryer is that a suspension is fed to the droplet generator and the droplets formed are mixed with hot air and dry as they fall down the drying tower. In the particle collector at the bottom of the setup, the dried particles are separated from the drying air. Droplet Generator. The droplet generator chosen for the droplet dryer is a slightly modified version of the commercial JetCutter type S available from GeniaLab GmbH, Germany. The principle of jet cutting is explained by Vorlop and Breford27 and sketched in Figure 3. A liquid jet is formed by forcing a feed through a nozzle with an orifice diameter of

Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009 3659

Figure 3. Droplet generation principle of the JetCutter.

Figure 2. Schematic representation of the droplet dryer which consists of an atomizing device and a drying tower. Table 1. Key Specifications of the Droplet Dryer tower height inner diameter liquid flow rate droplet size air flow rate air temperature

6.0 m 0.2 m ≈0.4 g/s 200-1500 µm 0.1 m/s 25-250 °C

100-300 µm. This jet is sliced into cylindrical pieces by a fastrotating cutting wheel (500-15000 rpm) equipped with 24-240 strings. The liquid surface tension reshapes the cylinders into spherical droplets shortly after formation. The droplet size may be altered by adjusting the jet velocity or the rotational speed of the cutting wheel. Specifically, the JetCutter installed on the droplet dryer is capable of producing droplets with a mean diameter in the range 200-1500 µm. Drying Tower. The droplets generated dry as they fall through the drying tower. The tower is equipped with 12 ports at different levels, thus allowing sampling of the droplets as they dry. The sampling ports are located at distances from 100 to 5600 mm from the air distributor. More ports are located at the top of the tower as the droplets experience the most rapid evaporation here. The sampling technique is described below. The droplets travel in a cocurrent plug flow of hot air. Plug flow is ensured by forcing the preheated air through a sand bed of height 100 mm in the air distributor at the top of the tower. The pressure drop over the sand bed is at least 100 mmH2O during all experiments. Although the velocity of the air stream is only 0.1 m/s, the amount of air is very large compared to the

amount of moisture evaporating from the drying droplets. Thus, the relative humidity in the air is negligible. To maintain a constant air temperature, the tower is insulated and three separate sections are coiled with thermostat controlled heating cables. The temperature may be checked by a specially designed thermocouple, enabling measurements at an arbitrary sampling port without causing significant disturbance to the tower air flow field. As mentioned above, in the particle collector at the bottom of the setup, the dried particles are separated from the drying air which is exhausted to the ventilation. The particle collector is mounted directly onto the setup but is easily removed for particle recovery. The particles are used for morphology analysis by optical and scanning electron microscopy (SEM). Sampling and Analysis. Droplets are caught through the ports during drying, using a special device26 which allows sampling without drawing ambient air into the tower. A small aluminum foil container is partly filled with low vapor pressure paraffin oil and fixed into the sampling device. During sampling, the droplets hit the paraffin oil, sink to the bottom because of a difference in density, and evaporation ceases. The sample is subsequently subjected to Karl Fischer titration as described by Skoog et al.28 for water mass fraction determination. Obtaining the mass fractions from samples taken at different ports yields a measure of the drying rate, i.e. the droplet water content as a function of the distance traveled in the drying tower is knownsexamples are given in the next section. The velocity of the droplets inside the drying tower may be estimated to 1.5-2 m/s (the terminal droplet velocity plus the velocity of the drying air). Sampling of droplets falling at this velocity is a difficult task. Inevitably, the experimental results are lightly scattered. From experience, particularly the measurements obtained from the top ports (ports situated less than 1000 mm from the droplet generator) display some scattering. At this point, the droplets still contain a high amount of water when caught and the surface tension between the droplets and the paraffin oil retards the sinking of the droplets into the paraffin oil. Thus, evaporation is not immediately stopped during sampling which is a source of error. Nevertheless, the trends of the results given below are always clear and thereby enable evaluation of the influence of different compounds on the drying kinetics. Experimental Results and Discussion Using the droplet dryer, several series of experiments with multicomponent suspension feeds have been conducted. Each series consists of four individual experiments differing only in the concentration of one component. Specifically, all experi-

3660 Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009 Table 2. Insolubles, Carbohydrates, and Inorganic Salts Used As Formulation Ingredients

rice starch TiO2 maltodextrin dextrin NaCl Na2SO4

trade name

supplier

d50a

remarks

REMY FG Kronos 2044 Glucidex IT21 AVEDEX W80 Suprasel

Alsiano Kronos Roquette Avebe Akzo Nobel Sulquisa

5.2 µm 0.39 µm soluble soluble soluble soluble

Can gelatinize DE 20-23b DE 9.5-12.5b food grade pregrinded

a 50% fractile of the primary particle size as measured by a Malvern Mastersizer employing laser diffraction. b Number of dextrose equivalents in the chain.

Figure 5. SEM pictures of dried rice starch particles containing various amounts of maltodextrin. Feed maltodextrin contents are given in the subfigure captions while all bar lengths are 100 µm.

Figure 4. Drying curves for maltodextrin containing rice starch suspensions. The legend indicates the maltodextrin content in weight percent.

ments are conducted using suspensions containing water, 42 wt % insoluble rice starch primary particles, and various amounts of a third component. The materials used in this work are given in Table 2. Further, unless explicitly stated below the droplet dryer settings are identical during the experiments, including a drying air temperature of 150 °C and an initial droplet size of approximately 275 µm. In the following, the drying kinetics are represented as the mass fraction of water evaporated wevap as a function of the distance L traveled in the drying tower. wevap is found from wevap ) 1 -

w(1 - w0) m )1m0 w0(1 - w)

(1)

where m0 and m are, respectively, the initial and the actual mass of water in a drying droplet. However, the values of m0 and m are immeasurable and instead wevap is calculated from the latter term of eq 1. Here w0 and w are, respectively, the water mass fraction of the feed and that found in the sample by Karl Fischer titration. Carbohydrates. Dextrins and maltodextrins are often used as excipients in the spray drying of proteins.29,30 To investigate the effect of maltodextrin on the drying kinetics and morphology formation during spray drying, a series of experiments has been conducted, changing only the maltodextrin content between experiments as described above. Figures 4 and 5 show the experimental measurements and SEM pictures of the dried particlesslines are inserted in Figure 4 to clarify the trends of the drying kinetics. Note, that one of the experiments is conducted using a feed containing only water and rice starch to serve as a reference for the other experiments. This experiment is simply referred to as the reference experiment throughout this paper. Figure 4 reveals that the drying rate is only significantly affected by the addition of maltodextrin at the highest content

Figure 6. Dried particles from suspensions containing only water and rice starch dried at various temperatures. The bars indicates a length of 100 µm .

level. Interestingly, the dried particles produced from the suspensions containing 3.24 and 6.45 wt % maltodextrin have suffered severe gelation; see Figure 5c and d. According to Spigno and De Faveri,31 the degree of gelation for rice starch increases with temperature and moisture content indicating that the particles mentioned have experienced elevated temperatures early in the drying process. This is supported by Figure 6 where suspensions containing only water and rice starch (≈ 24 wt %) have been dried at different air temperatures, leading to corresponding differences in the droplet temperature. For particles dried at 200 and 250 °C, there appears to be a layer of intact primary particles around the gelatinized region. This indicates that this layer has been dry before the temperature has risen to gelatinize a wet core. Evidently, the degree of gelation increases with an increase in droplet temperature dependent on the moisture content. Thus, Figures 4 and 5 indicate that addition of maltodextrin affects the droplet temperature during drying rather than the drying kinetics. A possible explanation is that the presence of maltodextrin reduces the droplet surface water activity and thereby the surface vapor pressure. This would be expected to lower the evaporation rate. However, a temperature rise almost compensates for the water activity loss and the vapor pressure is increased which gives an almost unaffected evaporation rate. Corresponding conclusions may be drawn from a series of experiments where various amounts of dextrin have been added

Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009 3661

Figure 7. Drying curves for rice starch suspensions containing dextrin as the third component. The legend indicates the dextrin content in weight percent.

Figure 8. SEM pictures of dried rice starch particles containing various amounts of dextrin. Feed dextrin contents are given in the subfigure captions while all bar lengths are 100 µm. The reference experiment is not included but shown in Figure 5a.

to the reference suspension containing water and rice starch. Figure 7 shows that the drying rate is only significantly influenced by the dextrin at the highest content level. Although the particles have suffered less gelation during drying, Figures 7 and 8 are very similar to the results for the maltodextrin. Morphology. The formation of the observed particle morphologies of the dried carbohydrate containing suspensions (Figures 5 and 8) may be explained by the theory of Charlesworth and Marshall5 although the theory is derived for droplets containing dissolved solids. According to the theory, if the solid crust formed around the droplet at a distinct point in time during drying (the beginning of stage C in Figure 1) is sufficiently porous, no change in morphology occurs during the falling rate period. The water is merely transported through the pores to the surface where it evaporates. Thus, a solid and dense structure may form as is the case for several of the particles in Figures 5 and 8. This course of drying is sketched in Figure 9 (sequence a) and is also discussed by Walker and Reed.32 Figure 9 shows the most likely sequences for formation of the particle morphologies observed throughout the present work. The figure is based on the work of several authors5,32,33 and extended so that the figure is valid for the drying of materials which may gelatinize. Contrary to the above, the crust formed around the drying droplet may become plastic and nonporous as a result of gelation. As the particle temperature increases, vapor forms and

a positive pressure is created inside the particle because the plastic crust is not sufficiently permeable. Consequently, a bubble forms within the particle and inflation might occur during the remaining part of the drying. This is also shown in Figure 9 (sequence c) and leads to the formation of hollow, gelatinized particles. A number of particles in Figures 5 and 8 are hollow, and it appears that especially particles dried from suspensions containing 6.45 wt % maltodextrin have experienced considerable inflation. In this case, the particle diameter is significantly larger than the initial droplet size of approximately 275 µm. Even further on the formation of hollow particles is that this process might proceed as shown in Figure 10. As described in the introduction, the droplet shrinks during the constant rate period but when the primary particles at the droplet surface have packed as closely as possible a solid phase forms around the droplet (Figure 10a). The droplet can shrink no further, and water is drawn to the surface through capillary action in the porous crust or, alternatively, a liquid/water interface is formed inside the crust. In both cases, liquid is transported from the droplet center, dragging along primary particles which are deposited inside the thickening crust. The lack of water and primary particles in the droplet center causes the formation of an internal void (Figure 10b). The void continues to grow while the crusts thickens until a dried hollow particle is formed (Figure 10c). The theory is supported by Crosby and Marshall,34 Lukasiewicz,33 Walker and Reed,32 and Minoshima et al.,35 and as shown in Figure 9 may lead to hollow gelatinized or nongelatinized particles. Inorganic Salt. The effects of reducing the water activity on the developments in drying rate and droplet temperature are further investigated by a series of experiments where various amounts of Na2SO4 are added to the reference suspension. Although there is considerable scattering of the measurements from the earliest phase of the drying process, Figure 10 illustrates that dissolved Na2SO4 has the same effect on the drying kinetics as the carbohydrates. Only a high initial content of the added component significantly lowers the drying rate. Further to this is that particles produced from Na2SO4 containing suspension are similar in morphology to the carbohydrate containing particlesslow Na2SO4 content forms a solid, dense morphology while high content forms hollow particles (examples are shown in Figure 12a and b). However, the hollow Na2SO4 containing particles are only slightly gelatinized. This means that on the one hand the particles have not been hot enough to gelatinize and on the other hand the temperature must have been sufficiently elevated to built an internal vapor pressure. The void may form as described above but a substantial internal pressure is necessary to avoid particle collapse. The surface of the particles appears to have poor permeability (Figure 12c) and thus provides the vapor transport resistance for internal pressure increase. An experiment where a high amount of NaCl is added to the reference suspension produces particles (Figure 12d) very similar to those from high carbohydrate content, i.e. gelatinized, hollow particles. Thus, it appears that adding an inorganic salt instead of a carbohydrate results in a similar course of drying. Although both the salts and the carbohydrates lower the surface water activity, the two classes of compounds could do this differently. The salts change the chemical potential while the carbohydrates might form a film with low water diffusivity at the droplet surface. Evans and Haisman36 report that the addition of Na2SO4 to starch suspensions raises the gelatinization temperature signifi-

3662 Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009

Figure 9. Sequences for formation of observed morphologies. The sequences are an extension of the theories outlined by Charlesworth and Marshall,5 Lukasiewicz,33 and Walker and Reed.32

Figure 10. Detailed sequence for formation of hollow particles. The darker and lighter colors indicate primary particles and water, respectively.

Figure 12. SEM pictures of dried rice starch particles containing various amounts of Na2SO4 or NaCl. Feed contents of the third component are given in the subfigure captions. Figure 11. Drying curves for rice starch suspensions containing Na2SO4 as the third component. The legend indicates the Na2SO4 content in weight percent.

cantly while the gelatinization temperature is less affected by the addition of NaCl. This may explain that only the NaCl particles are gelatinized. Further, the vapor pressure over a saturated NaCl solution is e.g. 9.1 kPa at 50 °C.37 The same vapor pressure over a saturated Na2SO4 solution is obtained at 47 °C38 while it is obtained at 44 °C for pure water.39 Thus, it is possible that the NaCl containing droplets have experienced the highest temperatures during drying, leading to enhanced gelation. Insolubles. The influence of adding insoluble TiO2 primary particles to the reference suspension on the drying behavior is shown in Figure 13 while the final particle morphology is given in Figure 14a, c, and e. Clearly, the drying rate is unaffected regardless of the amount of TiO2 added and the lack of rice starch gelation reveals that the particles have not experienced elevated temperatures. This is because TiO2, contrary to carbohydrates and inorganic salts, does not influence the water activity being insoluble in water and does not form a film with low water diffusivity.

Figure 13. Drying curves for rice starch suspensions containing insoluble TiO2 particles as the third component. The legend indicates the TiO2 content in weight percent.

The primary rice starch and TiO2 particles have average sizes of 5.2 and 0.39 µm, respectively, as given in Table 2. The presence of the smaller TiO2 primary particles might be expected to have an effect on the structure of the drying particle and thereby influence the internal water transport and consequently the drying kinetics during the falling rate period. However, if so, this appears to be of minor importance.

Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009 3663

Figure 16. Semiquantitative titanium concentration profiles of the particles shown in Figure 14. The subfigure captions give the amount of TiO2 in the feed suspension.

Figure 14. SEM pictures of various TiO2 containing particles are shown to the left where the dotted lines give the positions of the indentations. Corresponding EDX element maps of titanium are shown to the right, including areas chosen for titanium concentration profile determinations. The amount of TiO2 in the feed suspension is given in the subfigure captions while all bar lengths are 100 µm

The titanium is concentrated at the entire surface, including the indentation surface. To investigate how the surfaces may have become titanium covered, the element maps are converted into profiles using the image analysis tool available in Matlabsa commercial computer software package. Also using Matlab, the data are smoothed by statistical data processing. The profiles of Figure 16 are traced from a point chosen close to the particle center but precluding the indentations. The trace ends just beyond the particle surface as shown in Figure 14. Note that analyses of this kind are merely semiquantitative, and thus, no numerical values are given in Figure 16. A possible explanation of the over-representation of titanium at the particle surface might be that during drying, TiO2 primary particles are carried along with the liquid when it is transported toward the surface through the pores between the rice starch primary particles. This, however, seems unlikely considering the absence of a clear positive titanium concentration gradient in the profiles. More probable, the titanium accumulates at the surface during the constant rate period when the droplet shrinks. During shrinkage the rice starch primary particles recede but the TiO2 deposit at the surfacespossibly due to a filtering effect. Conclusion

Figure 15. Formation of an indentation based on the work of Bu¨ttiker.22

The sole effect of adding TiO2 is that the particles formed display large indentations. This may be explained by the theory discussed by Bu¨ttiker,22 Duffie and Marshall,40,41 and Shaw.42 The theory is sketched in Figure 15 as well as in Figure 9 (sequence b), and it is assumed that the droplet does not rotate during drying. The front of the droplet is subjected to a substantial drag, increasing the transfer of heat to the droplet and the transfer of mass (i.e., evaporated liquid) from the droplet. The evaporation from the rear surface is slower, drawing liquid to the front where evaporation is quicker. The flow of liquid carries along primary particles and the rear surface is drawn inward, forming an indentation. In the SEM pictures of Figure 14, the TiO2 primary particles are seemingly concentrated at the particle surfaces. This is verified by employing energy dispersive X-ray analysis (EDX) for element mapping of titanium (Figures 14b, d, and f).

The droplet drying process of multicomponent suspensions has been investigated experimentally. The experiments were conducted using a newly constructed apparatus which is designed for in situ drying experiments so that the droplet drying process resembles that of conventional spray drying. With focus on drying of formulations containing carrier materials and excipients relevant to the biotechnological and pharmaceutical industries, the drying kinetics have been measured and the particle morphology has been analyzed. The results show that adding compounds such as dextrin, maltodextrin, NaCl, or Na2SO4 to a suspension consisting of water and insoluble rice starch primary particles causes an increase in the droplet temperature rather than a reduction of the drying rate. The compounds added are believed to change the course of droplet drying by reducing the surface water activity by either changing the chemical potential (salts) or forming a film with low water diffusivity (carbohydrates). This should be considered when altering a formulation containing temperature sensitive active ingredients. An increased droplet temperature during drying was indicated by gelation of the rice starch.

3664 Ind. Eng. Chem. Res., Vol. 48, No. 7, 2009

No increase in droplet temperature nor change in drying rate was observed when insoluble TiO2 primary particles were added to the reference suspension of water and rice starch. This is explained by the fact that an insoluble compound does not affect the chemical potential or form a film with low water diffusivity. Observed morphologies include particles which are solid, have indentations, or are hollow. Sequences for formation of the observed morphologies are given and explained in detail to contribute to the ongoing effort in the literature to understand morphology formation processes. Acknowledgment The authors are grateful for the support of the Novozymes Bioprocess Academy and the CHEC Research Centre. Literature Cited (1) Nesic, S.; Vodnik, J. Kinetics of droplet evaporation. Chem. Eng. Sci. 1991, 46, 527–537. (2) Farid, M. A new approach to modelling of single droplet drying. Chem. Eng. Sci. 2003, 58, 2985–2993. (3) Sloth, J.; Kiil, S.; Jensen, A. D.; Andersen, S. K.; Jørgensen, K.; Schiffter, H.; Lee, G. Model Based Analysis of the Drying of a Single Solution Droplet in an Ultrasonic Levitator. Chem. Eng. Sci. 2006, 61, 2701– 2709. (4) El-Sayed, T. M.; Wallack, D. A.; King, C. J. Changes in particle morphology during drying of drops of carbohydrate solutions and food liquids - part 1: Effects of composition and drying conditions. Ind. Eng. Chem. Res. 1990, 29, 2346–2354. (5) Charlesworth, D. H.; Marshall, W. R. Evaporation from Drops Containing Dissolved Solids. AIChE J. 1960, 6, 9–23. (6) Malakhovskii, A. N. Drying Kinetics of Single Drops by Ferrite Suspensions. SoV. Powder Metall. Metal Ceram. 1980, 19, 76–78. (7) Sano, Y.; Keey, R. B. The Drying of a Spherical Particle Containing Colloidal Material Into a Hollow Sphere. Chem. Eng. Sci. 1982, 37, 881– 889. (8) Cheong, H. W.; Jeffreys, G. V.; Mumford, C. J. A Receding Interface Model for the Drying of Slurry Droplets. AIChE J. 1986, 32, 1334–1346. (9) Ranz, W. E.; Marshall, W. R. Evaporation from Drops - Part 1. Chem. Eng. Progress 1952, 48, 141–146. (10) Ranz, W. E.; Marshall, W. R. Evaporation from Drops - Part 2. Chem. Eng. Progress 1952, 48, 173–180. (11) Hecht, J. P.; King, J. Spray Drying: Influence of Developing Drop Morphology on Drying Rates and Retention of Volatile Substances. 1. Single-Drop Experiments. Ind. Eng. Chem. Res. 2000, 39, 1756–1765. (12) Masters, K. Spray Drying Handbook, 5th ed.; Longman Scientific and Technical: Harlow, U.K., 1991. (13) Lin, J.-C.; Gentry, J. W. Spray Drying Drop Morphology: Experimental Study. Aerosol Sci. Technol. 2003, 37, 15–32. (14) Toei, R.; Okazaki, M.; Furuta, T. Drying Mechanisms of a NonSupported Droplet. Proceedings of the First International Drying Symposium, Princeton, New Jersey, August 3–5, 1978; pp 53-58. (15) Groenewold, C.; Mo¨ser, C.; Groenewold, H.; Tsotsas, E. Determination of single-particle drying kinetics in an acoustic levitator. Chem. Eng. J. 2002, 86, 217–222. (16) Kastner, O.; Brenn, G.; Rensink, D.; Tropea, C. The Acoustic Tube Levitator - a Novel Device for Determining the Droplet Kinetics of Single Droplets. Chem. Eng. Technol. 2001, 24, 335–339. (17) Yarin, A. L.; Brenn, G.; Kastner, O.; Tropea, C. Drying of acoustically levitated droplets of liquid-solid suspensions: Evaporation and crust formation. Phys. Fluids 2002, 14, 2289–2298. (18) Toei, R.; Furuta, T. Drying of a droplet in a non-supported state. AIChE Symp. Ser. 1982, 78, 111–117. (19) Yarin, A. L.; Brenn, G.; Kastner, O.; Rensink, D.; Tropea, C. Evaporation of acoustically levitated droplets. J. Fluid Mech. 1999, 339, 151–204.

(20) Adhikari, B.; Howes, T.; Bhandari, B. R.; Truong, V. Experimental Studies and Kinetics of Single Drop Drying and their Relevance in Drying of Sugar-Rich Foods: A Review. Int. J. Food Prop. 2000, 3, 323–351. (21) Chen, X. D. Heat-Mass Transfer and Structure Formation During Drying of Single Food Droplets. Drying Technol. 2004, 22, 179–190. (22) Bu¨ttiker, R. Erzeugung Von Gleich Grossen Tropfen Mit Suspendiertem Und Geloestem Feststoff Und Deren Trocknung Im Freien Fall. Ph.D. thesis, Eidgeno¨ssische Technische Hochschule Zu¨rich, Switzerland, 1978. (23) Wallack, D. A.; El-Sayed, T. M.; King, C. J. Changes in particle morphology during drying of drops of carbohydrate solutions and food liquids - part 2: Effects on drying rate. Ind. Eng. Chem. Res. 1990, 29, 2354–2357. (24) Meerdink, G. Drying of Viscous, Shrinking Products: Modelling and Experimental Validation. Ph.D. thesis, Landbouwuniversiteit te Wageningen, The Netherlands, 1993. (25) Zbicinski, I. Development and experimental verification af momentum, heat and mass transfer model in spray drying. Chem. Eng. J. 1995, 58, 123–133. (26) Jørgensen, K. Drying Rate and Morphology of Slurry Droplets. Ph.D. thesis, Technical University of Denmark, 2005. (27) Vorlop, K.-D.; Breford, J. Verfahren und Vorrichtung Zur Herstellung Von Teilchen Aus Einem Flu¨ssige Medium. German Patent DE 4424998 A1, 1994. (28) Skoog, D. A.; West, D. M.; Holler, F. J. Fundamentals of Analytical Chemistry, 7th ed.; Saunders College Publishing: Orlando, FL, 1996; pp 381-383. (29) Millqvist-Fureby, A.; Malmsten, M.; Bergenståhl, B. Spray Drying of Trypsin - Surface Characterisation and Activity Preservation. Int. J. Pharm. 1999, 188, 243–253. (30) Selivanov, A. S. Stabilization of Cellulases Using Spray Drying. Eng. Life Sci. 2005, 5, 78–80. (31) Spigno, G.; De Faveri, D. M. Gelatinization Kinetics of Rice Starch Studied by Non-Isothermal Technique: Influence of Extraction Method, Water Concentration and Heating Rate. J. Food Eng. 2004, 62, 337–334. (32) Walker, W. J., Jr.; Reed, J. S. Influence of Slurry Parameters on the Characteristics of Spray-Dried Granules. J. Am. Ceram. Soc. 1999, 82, 1711–1719. (33) Lukasiewicz, S. J. Spray-Drying Ceramic Powders. J. Am. Ceram. Soc. 1989, 72, 617–624. (34) Crosby, E. J.; Marshall, W. R. Effects of drying conditions on the properties of spray-dried particles. Chem. Eng. Progress 1958, 54, 56–63. (35) Minoshima, H.; Matsushima, K.; Liang, H.; Shinohara, K. Basic Model of Spray Drying Granulation. J. Chem. Eng. Jpn. 2001, 34, 472– 478. (36) Evans, I. D.; Haisman, D. R. The Effect of Solutes on the Gelatinization Temperature Range of Potato Starch. Starch 1982, 34, 224– 231. (37) Apelblat, A.; Korin, E. The Vapour Pressures of Saturated Aqueous Solutions of Sodium Chloride, Sodium Bromide, Sodium Nitrate, Sodium Nitrite, Potassium Iodate, and Rubidium Chloride at Temperatures from 227 to 323 K. J. Chem. Thermodyn. 1998, 30, 59–71. (38) Apelblat, A.; Korin, E. The Vapour Pressure of Water over Saturated Solutions of Sodium Sulfate, Calcium Bromide, Ferric Chloride, Zinc Nitrate, Calcium Nitrate, and Lithium Nitrate at Temperatures from 278.15 to 323.15 K. J. Chem. Thermodyn. 2002, 34, 1621–1637. (39) CRC Handbook of Chemistry and Physics, 87th ed.; Lide, D. R., Ed.;Taylor and Francis: Boca Raton, FL, 2007; http:/www.hbcpnetbase.com. (40) Duffie, J. A.; Marshall, W. R. Factors Influencing the Propeties of Spray-Dried Materials. Chem. Eng. Progress 1953, 49, 417–423. (41) Duffie, J. A.; Marshall, W. R. Factors Influencing the Propeties of Spray-Dried Materials. 2. Drying Studies. Chem. Eng. Progress 1953, 49, 480–486. (42) Shaw, F. V. Spray Drying: A Traditional Process for Advanced Applications. Ceram. Bull. 1990, 69, 1484–1489.

ReceiVed for reView June 24, 2008 Accepted January 15, 2009 IE800983W